KFU. Card publication. Nonlinear Perturbation-based Non-Convex Optimization over Time-Varying Networks

KAZAN
FEDERAL UNIVERSITY

- RUS
- 中文
- ES

NONLINEAR PERTURBATION-BASED NON-CONVEX OPTIMIZATION OVER TIME-VARYING NETWORKS

Form of presentation

Articles in international journals and collections

Year of publication

2024

��

��

Authors, employees of KFU

Gabidullina Zulfiya Ravilevna, author

Other authors

Doostmohammadian Mohammadreza , author

Rabiee Hamid R. , author

Bibliographic description in the original language

Mohammadreza Doostmohammadian, Zulfiya R. Gabidullina, and Hamid R. Rabiee. Nonlinear Perturbation-based Non-Convex Optimization over Time-Varying Networks. arXiv preprint arXiv:2408.02269, 2024 Submitted on 5 Aug 2024 https://doi.org/10.48550/arXiv.2408.02269

Annotation

arxiv.org

Keywords

Distributed Algorithm, Non-Convex Local Optimization, Network Science, Perturbation Theory

The name of the journal

arxiv.org

URL

https://www.arxiv.org/abs/2408.02269

Please use this ID to quote from or refer to the card

https://repository.kpfu.ru/eng/?p_id=302892&p_lang=2

Full metadata record

Field DC	Value	Language
dc.contributor.author	Gabidullina Zulfiya Ravilevna	ru_RU
dc.contributor.author	Doostmohammadian Mohammadreza	ru_RU
dc.contributor.author	Rabiee Hamid R.	ru_RU
dc.date.accessioned	2024-01-01T00:00:00Z	ru_RU
dc.date.available	2024-01-01T00:00:00Z	ru_RU
dc.date.issued	2024	ru_RU
dc.identifier.citation	Mohammadreza Doostmohammadian, Zulfiya R. Gabidullina, and Hamid R. Rabiee. Nonlinear Perturbation-based Non-Convex Optimization over Time-Varying Networks. arXiv preprint arXiv:2408.02269, 2024 Submitted on 5 Aug 2024 https://doi.org/10.48550/arXiv.2408.02269	ru_RU
dc.identifier.uri	https://repository.kpfu.ru/eng/?p_id=302892&p_lang=2	ru_RU
dc.description.abstract	arxiv.org	ru_RU
dc.description.abstract	Decentralized optimization strategies are useful for a range of applications from networked estimation to distributed machine learning. This paper studies finite-sum minimization problems described over a network of nodes and proposes a computationally efficient algorithm that not only solves distributed convex problems but also optimally finds the solution to local non-convex objective functions. The algorithm is in single-timescale with no extra inner consensus loop and evaluates one gradient entry per node per time in contrast to batch gradient optimization in some literature. Further, the algorithm addresses link-level nonlinearity representing, for example, logarithmic quantization of the exchanged data. Leveraging perturbation-based theory along with algebraic Laplacian network analysis allows to prove optimal convergence and dynamics stability over time-varying and switching networks. The time-varying network setup might be due to packet drops and link failures. Despite the nonlinear nature of the dynamics, we prove exact convergence in the face of odd sign-preserving sector-bound nonlinear data transmission over the links. Illustrative numerical simulations further highlight our contributions.	ru_RU
dc.language.iso	ru	ru_RU
dc.subject	Distributed Algorithm	ru_RU
dc.subject	Non-Convex Local Optimization	ru_RU
dc.subject	Network Science	ru_RU
dc.subject	Perturbation Theory	ru_RU
dc.title	Nonlinear Perturbation-based Non-Convex Optimization over Time-Varying Networks	ru_RU
dc.type	Articles in international journals and collections	ru_RU